vcf not indexing

Question

[alex]

Trying to index vcf file but getting the following

tabix -p vcf dbsnp_138.hg19.vcf.gz
Not a BGZF file: dbsnp_138.hg19.vcf.gz
tbx_index_build failed: dbsnp_138.hg19.vcf.gz

Thoughts on how to proceed? Thanks!

Answer 1

[Sean Davis]

Looks to me like the dbsnp file is not bgzipped?

gunzip dbsnp_138.hg19.vcf.gz
bgzip dbsnp_138.hg19.vcf
tabix -p vcf dbsnp_138.hg19.vcf.gz

[jespinoz]

@Sean Davis you just saved me a lot of frustration. I found this after a few searches and it worked. THanks

[marongiu.luigi]

I had the same problem: when I compressed with gunzip <file>.vcf tabix gave the error: tbx_index_build failed:<file>.vcf.gz but it worked with `bgzip <file>.vcf.

[ATpoint]

Of course it does. bgzip does blockwise (therefore the b in bgzip) compression of the file, which tabix relies on. That enables tabix to quickly retrieve data by only very partially decompressing a (sometimes hugh) file, guided by the index. gzip does not do blockwise compression.